1 /*- 2 * Copyright (c) 2005-2014 Rich Felker, et al. 3 * 4 * Permission is hereby granted, free of charge, to any person obtaining 5 * a copy of this software and associated documentation files (the 6 * "Software"), to deal in the Software without restriction, including 7 * without limitation the rights to use, copy, modify, merge, publish, 8 * distribute, sublicense, and/or sell copies of the Software, and to 9 * permit persons to whom the Software is furnished to do so, subject to 10 * the following conditions: 11 * 12 * The above copyright notice and this permission notice shall be 13 * included in all copies or substantial portions of the Software. 14 * 15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 16 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 17 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. 18 * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY 19 * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, 20 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE 21 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. 22 */ 23 #include <sys/cdefs.h> 24 #if defined(LIBC_SCCS) && !defined(lint) 25 #if 0 26 __FBSDID("$FreeBSD: head/lib/libc/string/memmem.c 315468 2017-03-18 00:53:24Z emaste $"); 27 #else 28 __RCSID("$NetBSD: memmem.c,v 1.3 2018/10/15 19:32:48 christos Exp $"); 29 #endif 30 #endif /* LIBC_SCCS and not lint */ 31 32 #if !defined(_KERNEL) && !defined(_STANDALONE) 33 #include <string.h> 34 #include <stdint.h> 35 #else 36 #include <lib/libkern/libkern.h> 37 #endif 38 39 static char *twobyte_memmem(const unsigned char *h, size_t k, 40 const unsigned char *n) 41 { 42 uint16_t nw = n[0] << 8 | n[1], hw = h[0] << 8 | h[1]; 43 for (h += 2, k -= 2; k; k--, hw = hw << 8 | *h++) 44 if (hw == nw) return __UNCONST(h - 2); 45 return hw == nw ? __UNCONST(h - 2) : 0; 46 } 47 48 static char *threebyte_memmem(const unsigned char *h, size_t k, 49 const unsigned char *n) 50 { 51 uint32_t nw = n[0] << 24 | n[1] << 16 | n[2] << 8; 52 uint32_t hw = h[0] << 24 | h[1] << 16 | h[2] << 8; 53 for (h += 3, k -= 3; k; k--, hw = (hw|*h++) << 8) 54 if (hw == nw) return __UNCONST(h - 3); 55 return hw == nw ? __UNCONST(h - 3) : 0; 56 } 57 58 static char *fourbyte_memmem(const unsigned char *h, size_t k, 59 const unsigned char *n) 60 { 61 uint32_t nw = n[0] << 24 | n[1] << 16 | n[2] << 8 | n[3]; 62 uint32_t hw = h[0] << 24 | h[1] << 16 | h[2] << 8 | h[3]; 63 for (h += 4, k -= 4; k; k--, hw = hw << 8 | *h++) 64 if (hw == nw) return __UNCONST(h - 4); 65 return hw == nw ? __UNCONST(h - 4) : 0; 66 return 0; 67 } 68 69 #define MAX(a,b) ((a)>(b)?(a):(b)) 70 #define MIN(a,b) ((a)<(b)?(a):(b)) 71 72 #define BITOP(a,b,op) \ 73 ((a)[(size_t)(b)/(8*sizeof *(a))] op (size_t)1<<((size_t)(b)%(8*sizeof *(a)))) 74 75 /* 76 * Two Way string search algorithm, with a bad shift table applied to the last 77 * byte of the window. A bit array marks which entries in the shift table are 78 * initialized to avoid fully initializing a 1kb/2kb table. 79 * 80 * Reference: CROCHEMORE M., PERRIN D., 1991, Two-way string-matching, 81 * Journal of the ACM 38(3):651-675 82 */ 83 static char *twoway_memmem(const unsigned char *h, const unsigned char *z, const unsigned char *n, size_t l) 84 { 85 size_t i, ip, jp, k, p, ms, p0, mem, mem0; 86 size_t byteset[32 / sizeof(size_t)] = { 0 }; 87 size_t shift[256]; 88 89 /* Computing length of needle and fill shift table */ 90 for (i=0; i<l; i++) 91 BITOP(byteset, n[i], |=), shift[n[i]] = i+1; 92 93 /* Compute maximal suffix */ 94 ip = (size_t)-1; jp = 0; k = p = 1; 95 while (jp+k<l) { 96 if (n[ip+k] == n[jp+k]) { 97 if (k == p) { 98 jp += p; 99 k = 1; 100 } else k++; 101 } else if (n[ip+k] > n[jp+k]) { 102 jp += k; 103 k = 1; 104 p = jp - ip; 105 } else { 106 ip = jp++; 107 k = p = 1; 108 } 109 } 110 ms = ip; 111 p0 = p; 112 113 /* And with the opposite comparison */ 114 ip = (size_t)-1; jp = 0; k = p = 1; 115 while (jp+k<l) { 116 if (n[ip+k] == n[jp+k]) { 117 if (k == p) { 118 jp += p; 119 k = 1; 120 } else k++; 121 } else if (n[ip+k] < n[jp+k]) { 122 jp += k; 123 k = 1; 124 p = jp - ip; 125 } else { 126 ip = jp++; 127 k = p = 1; 128 } 129 } 130 if (ip+1 > ms+1) ms = ip; 131 else p = p0; 132 133 /* Periodic needle? */ 134 if (memcmp(n, n+p, ms+1)) { 135 mem0 = 0; 136 p = MAX(ms, l-ms-1) + 1; 137 } else mem0 = l-p; 138 mem = 0; 139 140 /* Search loop */ 141 for (;;) { 142 /* If remainder of haystack is shorter than needle, done */ 143 if ((size_t)(z-h) < l) return 0; 144 145 /* Check last byte first; advance by shift on mismatch */ 146 if (BITOP(byteset, h[l-1], &)) { 147 k = l-shift[h[l-1]]; 148 if (k) { 149 if (mem0 && mem && k < p) k = l-p; 150 h += k; 151 mem = 0; 152 continue; 153 } 154 } else { 155 h += l; 156 mem = 0; 157 continue; 158 } 159 160 /* Compare right half */ 161 for (k=MAX(ms+1,mem); k<l && n[k] == h[k]; k++); 162 if (k < l) { 163 h += k-ms; 164 mem = 0; 165 continue; 166 } 167 /* Compare left half */ 168 for (k=ms+1; k>mem && n[k-1] == h[k-1]; k--); 169 if (k <= mem) return __UNCONST(h); 170 h += p; 171 mem = mem0; 172 } 173 } 174 175 void *memmem(const void *h0, size_t k, const void *n0, size_t l) 176 { 177 const unsigned char *h = h0, *n = n0; 178 179 /* Return immediately on empty needle */ 180 if (!l) return __UNCONST(h); 181 182 /* Return immediately when needle is longer than haystack */ 183 if (k<l) return 0; 184 185 /* Use faster algorithms for short needles */ 186 h = memchr(h0, *n, k); 187 if (!h || l==1) return __UNCONST(h); 188 k -= h - (const unsigned char *)h0; 189 if (k<l) return 0; 190 if (l==2) return twobyte_memmem(h, k, n); 191 if (l==3) return threebyte_memmem(h, k, n); 192 if (l==4) return fourbyte_memmem(h, k, n); 193 194 return twoway_memmem(h, h+k, n, l); 195 } 196